INDEX

Explanations

phrases that signal actions or processes involving objectives and outcomes

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/1-res-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.1.hook_resid_post

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

msgstr

-1.88

enda

-1.69

lier

-1.60

Ãªncia

-1.59

atile

-1.51

rehens

-1.50

"}](#

-1.47

 msgid

-1.44

illac

-1.42

 tricks

-1.42

POSITIVE LOGITS

«

2.07

¬

1.85

 hers

1.73

asts

1.70

 yours

1.69

his

1.65

Īĺ

1.62

ķ

1.60

uate

1.59

Ĥ

1.59

Activations Density 0.249%

phrases that signal actions or processes involving objectives and outcomes

No Comments

No Known Activations